Generative AI tools, such as ChatGPT, make it easier for publishers to conduct customer research, generate product ideas and plans, and analyse feedback.

In 2022, U.K.-based Financial Times searched for ways to grow its subscriber base beyond the professional readers, its core audience motivated by a desire to succeed professionally.

After rigorous research, FT designed from scratch a new app, The Edit, aimed at time-strapped young adults who wish to understand complex global issues but feel overwhelmed by endless news feeds.

Let’s hire AI to recreate this product discovery (the Financial Times’ team was not involved in this experiment):

Prompt: Act as … a product manager for the Financial Times. Plan a research project aimed at segmenting audiences based on their preference for breaking and in-depth news as well as willingness to pay for online news. 

Hint: To get the most out of an AI bot, focus on a specific topic and use language that relates to it. Start with a clear question to guide the conversation. You can also make the bot pretend to be someone specific and respond from their perspective.

Example response from ChatGPT: “As a product manager for the Financial Times, I would plan the following research project:

1. Develop a survey that includes questions on participants’ reading habits, their preference for breaking and in-depth news, and their willingness to pay for online news. The survey can also include demographic information such as age, income, and education level.
2. Recruit a diverse sample of participants from our current audience, as well as potential new audiences. This can be done through social media, e-mail campaigns, or by working with market research firms.
3. Once data is collected, use statistical techniques such as cluster analysis or decision trees to segment the audience based on their preference for breaking and in-depth news and their willingness to pay for online news.
4. Identify the segments that have similar characteristics and preferences, and assign them a name that reflects their characteristics, such as “Breaking News Seekers” or “In-Depth News Enthusiasts.”
5. Validate the segments by conducting qualitative research such as in-depth interviews or focus groups to understand the characteristics of each segment more in-depth and ensure that the segments are accurate and well-defined.
6. Develop a strategy to target each segment based on their preferences and willingness to pay for online news. This can include creating different subscription packages, tailored content, and targeted marketing campaigns.
7. Implement the strategy and monitor the results over time by tracking key metrics such as conversion rates, retention rates, and customer lifetime value for each segment.”

Prompt: Develop ... a survey that includes questions on participants’ reading habits, their preference for breaking and in-depth news, and their willingness to pay for online news. 

Hint: Follow-up on the conversation and ask AI to draft research tools it mentioned, such as a survey on reading habits or a scenario for in-depth interviews or focus groups. The AI remembers the previous conversation, and it might refer to it.

In the following example response, AI comes up with a 12-question survey. Apart from the questions suggested in the prompt, new ones were added, for example, about the importance of a source’s trustworthiness.

Example response from ChatGPT: “(...)

2. What type of news do you prefer reading?

• Breaking news
• In-depth analysis and features
• Both equally

3. Are you willing to pay for online news?

• Yes
• No
• I currently pay for online news

(...)

9. How important is the trustworthiness of the news source to you?

• Very important
• Somewhat important
• Not important”

Prompt: Revise … the survey to include questions about readers’ experience with news, including gains or motivations to read it, and pains or frustrations when reading it. 

Hint: If you are not satisfied with a draft, simply ask ChatGPT to revise it. Suggest areas for further inquiry. For example, tell AI to use human-centred design frameworks such as gains and pains experienced by a customer when using a product or a service. 

Refer to established academic theories such as media and gratification theory suggesting that people have different personal motivations for engaging with media. You don’t need to train the AI about the established theories or frameworks, as it likely knows them already.

Example response from ChatGPT: “13. What are your personal motivations for reading the news?

• To stay informed about current events
• To pass time
• To make sense of the world
• To be entertained
• To feel like an active citizen
• To identify with a specific group or social identity
• Others (please specify)

14. What are the main pain points or frustrations you experience when reading the news?

• Lack of trustworthiness
• Bias or one-sidedness
• Too much sensationalism
• Overwhelming amount of information
• Lack of relevance to my interests
• Difficulty in finding credible sources
• Difficulty in separating facts from opinions”

Prompt: Empathise … Act as a news consumer who is frustrated by being overwhelmed by the amount of information and describe your experience. 

Hint: AI excels at role playing. Before talking to humans, you can chat with a machine pretending to be the person of your interest. The bot is capable of assuming the new role immediately and eagerly discusses the character’s experience, including her feelings. 

It can also answer further questions chat-style. For example, it may provide more details or examples similar to how a real-life, in-depth interview would look. The bot can also describe the character itself, drafting a persona representing the segment of the audience that you study. This is useful for setting up further conversations or feeding your marketing brief. It can map an example customer journey and discuss the role of each stage or touchpoint.

Obviously, machines do not subscribe to news, so use it as a playground and preparation for the real-life interviews with humans. Again, the bot can help you draft and test scenarios for such interviews and summarise them afterwards.

Example response from ChatGPT: “As a news consumer, I am constantly bombarded with a never-ending stream of information. It feels like there is always something new happening and I am struggling to keep up. I feel overwhelmed by the sheer volume of news and it’s hard for me to filter out what is important and relevant to me.

I am also confused by the conflicting information being reported by different sources. I am anxious that I am missing out on important information and that I am not well-informed as a result.

I am exhausted from consuming a large amount of news on a regular basis. It’s mentally and emotionally taxing. I am also distrustful of many sources of news, it’s hard for me to know which sources are reliable and which are not. (...)”

Prompt: Prototype … a new product that addresses these pains aimed at readers who value in-depth news more than breaking news and are willing to pay for online news. Describe the value proposition.

Hint: Continuing our design thinking exercise, after empathising with customers, you can ideate and prototype a solution to their pains. Again, AI is answering with a mix of bullet points linked to the previous conversation, meeting the set constraints with new ideas such as a money-back guarantee or a proposal to make the product advertising-free. 

Example response: “The product I am proposing is called ‘Insight News’. Insight News is a subscription-based online news platform that prioritises in-depth analysis, investigative reporting, and long-form journalism.

Value proposition:

• In-depth coverage: Insight News offers in-depth coverage and analysis of current events, giving readers a deeper understanding of the issues at hand.

• Quality over quantity: Insight News focuses on providing a select number of well-researched and expertly written articles each day, rather than a constant stream of breaking news. (...)

• Ad-free: Insight News is ad-free, providing readers with an uninterrupted reading experience.

• Satisfaction guarantee: Insight News offers a satisfaction guarantee, so that if the reader is not satisfied with the product, they can cancel their subscription and receive a full refund.”

Prompt: Ideate … better names for this product than “Insight News” and make clear the connection to the FT brand.

Hint: AI can assist you in brainstorming, offering endless suggestions based on your directions and feedback. For example, you can ask it to emphasise the nature of journalism delivered by the new product, the role of the newsroom, or the desired benefit the potential reader seeks.

In a few minutes, AI generated dozens of ideas, including one that sounded close to the name of the real product, “FT Edit,” that the Financial Times eventually launched.

Example responses from ChatGPT: “FT Insight”... “FT Deep Dive”... “FT Context”... “FT Curated”... “FT Editor’s Choice”... “FT Edits”... “FT Handpicked”... “FT Clear View”... “FT Uncover”... “FT Unbiased” (...)

Found other uses for generative AI in news marketing? E-mail me at greg.piechota@inma.org.

Greg’s Readers First newsletter is a public face of a revenue and media subscriptions initiative by INMA, outlined here. Subscribe here.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .